Introduction to Templates

Transcription

1 Introduction to Templates Module 1 of the Reference Data Readiness Course Julian M.N. Bourne JORD 2014 PCA and Fiatech

2 Module & Course Introduction Introduction to Templates is the first module of the Reference Data Readiness Course The Reference Data Readiness course helps us to: consistently analyze and communicate reference data needs produce discipline-specific reference data solutions collaborate on abstract reference data solutions This first module, Introduction to Templates helps us to: express the rules and meaning of existing data flows

3 Topics in this Module The Need for Reference Data Implicit Information Naming the Parts of Data The Shape of Information The Rules of Information Identifiers and Scope Capturing Information

4 Topic: The Need for Reference Data and templates, and iring and ISO The Introduction to Templates module will help describe existing data flows: But why is that even important? And how does that relate to reference data, templates, iring and ISO 15926? Understanding the need for reference data is the key

5 Engineering Data Today Is Like a Bucket of Nuts and Bolts Different Bolts: thread type & size cap type & size threaded length body length material strength condition etc. Different Data: format character encoding schema value type allowable range identifier registry context etc. Worse for data: the relationship between values is often not recorded at all. Its not just like a bucket of nuts and bolts, its like that bucket before parts became interchangeable and no two threads were even the same. Adapted from work Bucket of bolts under Creative Commons Attribution 3.0 Unported license attributed to Vinvent Racaniello

6 Reference Data and Interchangeability Part interchangeability revolutionized engineering: make parts to one specification then a part can be swapped for another enabled mass production reduced maintenance costs Reference data is part interchangeability for data: make data to one specification then a data source or target can be changed in a flow enables mass data reuse reduces integration costs but we re not quite there yet.

7 Identification Marks in Engineering For example A4-70 Aids in identification by asserting: thread type & geometry (metric) material (corrosion resistant stainless steel) tensile strength (annealed to 700MPa) Helps interpret more obvious properties like: diameter (10mm) body length (40mm) thread pitch (coarse) cap shape (17mm hex) Reduces knowledge and time required to use. A4-70 DIN ES ISO :1997 thread type: metric; material: Fe,Cr,Mo,Ni min. tensile strength: 700N/mm^2; etc.

8 Reference Data Stamps Meaning on Data applies the same principle to data stamps the data with a mark the mark points to reference data the reference data provides the meaning Example: Max. Press.: 80 maximum pressure? in what units? for the equipment design? for the selected equipment? for the operational process? for an alarm trigger? We might already have the data Reference data would provide clear answers to all these questions but its reference data that will preserve what that data means.

9 Reference Data Provides Data With Meaning Reference data adds meaning to data: data + reference data > data data + reference data = information Information that can then: be re-used be re-purposed cross between companies cross between work processes cross between disciplines... at low cost. Because you don t need to sort through a bucket of unlabelled nuts and bolts, because you don t need to guess about the meaning.

10 Connecting Reference Data to iring and ISO ISO a way of describing changing engineering data iring a community of people that develop ISO solutions Templates a kind of shorthand for ISO 15926, described in part 7 Templates and ISO are the how: discipline framework Reference data is the what: tight engineering-centric descriptions of data concepts The iring community: defines reference data collaborative cover whole industry needs without favoring competitors

11 Reference Data and the Introduction to Templates Reference data is: needed to preserve the meaning of data stamps the meaning directly on the data critical to ISO built around templates by the iring community The rest of this module: starting point for templates how to describe existing data using templates author use cases provide sample data Valuable inputs to making reference data.

12 Topic: Implicit Information using a spreadsheet example Column Column Label To stakeholder representatives: HELP NEEDED We put out the call for real-world content on we really need real world sample data to put on this page and to use for the course. Header Row A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press P Heading Cell Row Number Value

13 Columns In a spreadsheet: columns have a label (eg. A, B, C) and often a heading in row 1 (eg. Tag) the rest of the cells can contain values and most importantly: those values are usually a similar kind of thing the heading gives a clue as to what that kind of thing is We will use those last two features a lot in this course.

14 Rows In a spreadsheet: rows have a number (1, 2, 3, 4 etc.) the first row often isn t data, but a header the rest of the cells can contain values usually not past the last column heading though empty rows are not meaningful and most importantly: those values on the same row are related to the same thing and that thing is usually referred to in column A quite often, the values in column A cannot repeat sometimes, a combination of values cannot repeat

15 What is Implicit Information? Any notion: not explicitly mentioned in the spreadsheet; and necessary to populate the spreadsheet properly; or necessary to interpret the spreadsheet properly. Includes a lot of company and industry conventions. Our aim: make all information explicit So someone outside of: the workflow the discipline the company Can evaluate what is needed to: populate change interpret the spreadsheet.

16 Why is implicit information important? With it we can create a profile of the data: record real-world sample data record the rules that apply to the data record the nature of those rules This profile makes it possible to: re-use the workflow against different data re-purpose the same data in downstream applications provide high value inputs into reference data definition helping engineering companies avoid duplicate work

17 Working Towards Discovering Implicit Information return to the spreadsheet examine the obvious features columns especially ask some simple questions about the content and structure list some kinds of rules that can apply to content

18 Obvious Assumptions Column A: the values must all be tags the values must not repeat probably a functional location maybe physical equipment Columns B and C: the values must all be numbers some assumed unit of measure represent a kind of pressure A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press P

19 How to discover implicit information: Columns Look at the values in a column collectively and consider: Is the value a unique name or number for something (eg. tag in A)? Does that value refer to a physical thing? Could some other physical thing keep the same value and replace it? Is it a place where we put a physical thing? Is that a single real world place, or is it a place in a general design? Does the whole row describe what things can go in that place? What group of names or numbers should it be unique in? The spreadsheet? A design? A plant? A company? A 1 Tag 2 3 P

20 How to discover implicit information: Values If the value isn t a unique name or number Is the value some other kind of number? Is it numbering things in order (an ordinal number or index)? Is it counting quantities of things (a cardinal number or count)? Is the value a measurement (such as a pressure in B or C): What kind of measure is it (pressure, length, volume, power )? What unit of measure does it imply (bargs, kpa; m, ft; m^3; W etc.)? What is it measuring (maximum pressure, height, power draw etc.)? What other column refers to the thing being measured or counted (eg. A)? If the value isn t a measurement, number or unique name: Is the value a comment or description? What column refers to the thing being described, or is it the whole row? B 1 Nom. Press.

21 How to discover implicit information: Rows What about rows, is the order of rows significant in any way? which columns show the order progressing? is the order for the whole spreadsheet? or just for rows with other columns with matching values? which columns do they depend on? is that something that you can manage with a spreadsheet function? or does information outside of the spreadsheet govern the order?

22 How to discover implicit information: Other Considerations Are some values not allowed to repeat? In one column? In combinations of columns? Is two identical rows legal? Are there special values? Text in a number field? Does zero or -1 have a special meaning? Multiple numbers in one cell? Can the value in a cell be empty? What must be in the spreadsheet? What does it mean for a value in Eg. all equipment tags in plant a cell to be empty? XYZ? Does empty mean: Don t know? Eg. all pressure transponder No data? Not needed? Or something else? functional locations in some refinery stage? Questions to ask and answer about your data.

23 Topic: Naming the Parts of Data Putting names on the things we ve discovered: The things that columns represent. The way that values refer to those things. The things that rows represent. The thing that pulls all of that together. Then we name some common kinds of rules. Before naming though, lets work through a short, visual example.

24 Categorize Each Column As Completely As Possible A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 Transponder Functional Location in Splitter Z1 draw a box around the column label draw an oval somewhere convenient connect them with a line describe what a column value represents put that description in the oval Transponder Functional Location in Splitter Z identified by Tag A. describe how the value refers to that start a sentence with that description underneath

25 Extend the Sentence at the Bottom for Each Column Transponder Functional Location in Splitter Z1 Gauge Pressure in bars same process: box, line, oval, describe and extend sentence A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 Transponder Functional Location in Splitter Z identified by Tag A has nominal pressure B measured as gauge pressure in bars.

26 Join Like Things Together and Continue to Extend Sentence Transponder Functional Location in Splitter Z1 Gauge Pressure in bars similar process: but join like things together and extend sentence A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 Transponder Functional Location in Splitter Z identified by Tag A has: nominal pressure B measured as gauge pressure in bars; and maximum pressure C measured as gauge pressure in bars.

27 Naming the Ovals and Lines entity Transponder Functional Location in Splitter Z1 the oval is called an entity the line is called a role role A B C 1 Tag Nom. Press. Max. Press. 2

28 Naming the whole thing... relation description A B C 1 Tag Nom. Press. Max. Press. 2 role Transponder Functional Location in Splitter Z1 role Gauge Pressure in bars role the chain of boxes is called a relation the text is a description for the relation the whole thing: roles + relation + description is a template definition Transponder Functional Location in Splitter Z identified by Tag A has: nominal pressure B measured as gauge pressure in bars; and maximum pressure C measured as gauge pressure in bars.

29 Constraints: Naming the Rules Most rules in general: Collectively called Constraints. Formally, very complicated. Restrict to common, simple ones for our usage. The rules we will cover: This thing has to be in this spreadsheet. This value can only appear once. This combination of values can only appear once.

30 When Something Must Appear mandatory constraint Transponder Functional Location in Splitter Z1 Gauge Pressure in bars In some column, if an entity MUST appear: where the line starts on the oval place a filled dot This is called a mandatory constraint. A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 Transponder Functional Location in Splitter Z identified by Tag A has: nominal pressure B measured as gauge pressure in bars; and maximum pressure C measured as gauge pressure in bars. all transponder funclocs in splitter Z1 must be in this spreadsheet.

31 When A Value Can Only Appear Once uniqueness constraint Transponder Functional Location in Splitter Z1 Gauge Pressure in bars In some column, if a value can only appear once: where the role line ends above the box draw a line the length of the box This is called a uniqueness constraint. A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 Transponder Functional Location in Splitter Z identified by Tag A has: nominal pressure B measured as gauge pressure in bars; and maximum pressure C measured as gauge pressure in bars. a transponder can only have one nominal and one maximum pressure.

32 And On A Combination of Values uniqueness constraint Electrical Bus Equipment The same principle can be applied to multiple columns: extend the bar in-line Current in Amps A B C D E 1 Bus Equipment Current Wire Gauge Insulation 2 Electrical Bus identified by Tag A connected to equipment identified by Tag B drawing current C Amps. a bus can only connect once to each equipment.

33 Even Works with Intervening Columns Electrical Bus Current in Amps Just skip over non-relevant intervening columns: keep lines aligned gray dashes to link uniqueness constraint Equipment A B C D E 1 Bus Current Equipment Wire Gauge Insulation 2 Electrical Bus identified by Tag A connected to equipment identified by Tag C drawing current B Amps. a bus can only connect once to each equipment.

34 Even Works with Multiple Constraints Electrical Bus Current in Amps Stack the different constraints: so they don t line up gray dashes to link different uniqueness constraints A B C D E 1 Bus Current Equipment Wire Gauge Insulation 2 Equipment Insulation Coloring Electrical Bus identified by Tag A connected to equipment identified by Tag C drawing current B Amps. insulation colors have to be different for anything on the same bus.

35 Review Visual and Textual Vocabulary: ovals for entities boxes chained together for relations lines from ovals to boxes for roles dots for mandatory constraints (on the oval end) free lines for uniqueness constraints (stacked above boxes) Important: if you ever need to guess, just write it in text and leave it off the diagram but always record as much as you can however you can.

36 Topic: The Shape of Data If we reduce columns to square boxes Transponder Functional Location in Splitter Z1 Gauge Pressure in bars the diagram is similar to Object-Role Modelling (ORM). Add anything abnormal to ORM in grey, like the uniqueness constraint dashes and data modellers can still understand you. Transponder Functional Location in Splitter Z1 Gauge Pressure in bars

37 Tying Back to the Spreadsheet How does it link to your spreadsheet? Add spreadsheet column labels in grey (since they re not conventional) inside the box that corresponds to the column. Transponder Functional Location in Splitter Z1 A B C Gauge Pressure in bars This greatly helps data modellers too: when they start splitting up the relations and moving things around it keeps it linked back to your data.

38 Optional Values What about when a cell can be empty? Templates don t like empty. ORM doesn t like empty. Data modellers don t like empty. Transponder Functional Location in Splitter Z1 Gauge Pressure in bars A B C D Construction Plan Label Why don t they like empty? Because empty can have more than one meaning. Not applicable? Not specified? Missing? To be done (by whom?)? What if the construction plan label can be empty?

39 Showing Optional Values Add an open circle in grey (since it s not conventional) centered on where the role joins with the box and document why! (very important) Transponder Functional Location in Splitter Z1 Gauge Pressure in bars A B C D Construction Plan Label Communicates that in the spreadsheet the value is optional. Indicates to data modellers that this needs refactoring. The construction plan label is optional (document why, data modellers will fix it)

40 Documenting Optional Values Document by considering: What does empty mean? Can it always be empty? Or is it dependent on other values? Does it have to be empty in certain cases? Why were these choices made? an empty cell can hold an awful lot of implicit information, an awful lot of meaning Example: The construction plan label D must not be empty if the functional location is not subordinate to another functional location without a plan label. It may otherwise be empty to indicate that it is at the same plan location as its enclosing functional location. It must be empty on integral equipment subordinate functions, so that the plan location only needs to be changed in one place if the equipment is moved on the plan.

41 More On the Shape of Data Transponder Functional Location in Splitter Z1 Gauge Pressure in bars A B C D Transponder Functional Location in Splitter Z1 A A B C D Gauge Pressure in bars Construction Plan Label Construction Plan Label how data modellers think of it. (you don t need to learn this for this module)

42 Topic: Identifiers and Scope What is an identifier? a name (or number) that stands for something (other than itself) and is unique (within some specific list of things) Examples: equipment tags (within a plant scope ) serial numbers (within a make and model scope ) model numbers (within a make scope )

43 Purpose of Identifiers and Scope Identifiers: critical to the interoperability of data unique within a scope eg. Tag P83945 in Plant XYZ used so that we can make definitive statements about one thing or another and repurpose those statements in other situations Scope: the same identifier in a scope can refer to only one thing needs an identifier of its own eg. Equip. Tags in Plant XYZ Sometimes permitted: multiple identifiers for the same thing though usually discouraged in the same scope

44 Implicit Information: Format and Registry Needed to manage identifiers outside of origin tools. Format: the textual form eg. P83945 might have to start with a letter and be followed by five digits those rules are parts of its format. Registry: where/how to register/create new identifiers in some scope. Things to remember to record about format: can contain derived elements (check digits) sometimes ignores spaces, dashes, dots (P might be the same) information bearing identifiers (next slide)

45 Implicit: Information Bearing Identifiers Example: transponders in plant XYZ might all have an identifier like P83945 starts with a letter, followed by 5 digits but at plant XYZ, two things are special: P stands for pressure transponders, like P83945 T stands for temperature transponders, maybe like T84101 The first digit (8 in this case) stands for splitter Z1 Very important to record :: sometimes it might be the only data linkage A B C D E 1 Tag Nom. Press. Max. Press. Alarm Press. Shut. Press 2 T P C, not barg

46 Ramifications of Information Bearing Identifiers Transponder Functional Location in Splitter Z1 Gauge Pressure in bars or Temperature in Celsius data modellers shudder A B C might involve creating some weird entities (like an entity that is either a pressure or temperature)

47 How A Modeller Might See That Instead Changes: A appears 6 times B appears 3 times C appears 3 times 9 roles not on sheet Explicit key (Tag) for A Relations split B & C Role names added Entities wrap relations Arrows joining entities Dashed entity (data type) Other things could be added: constant references (to bar and Celsius) stop something being both temp. & pressure ORM Diagram (you don t have to learn this now)

48 Real World Data Topic: Capturing Information Entire Goal: how to create & use solid reference data Course Goal: how to create reference data Module Goal: how to describe information flows Describe Information Flows Analyse Flows Across Discipline The Importance of Real World Data Quick Checklist Template Breakdown Summary Create Discipline-Specific Reference Data Analyse Reference Data Across Disciplines Create Whole-Life-Cycle Reference Data

49 Importance of Real-World Sample Data Real-World Sample Data: Is essential to describe an information flow accurately. Is critical to the definition of reference data. Analyse: Always start with real-world sample data. Describe it using the techniques shown. Diagram it to check and complete. Release: description diagram sample data legal terms

50 Quick Checklist Write about the spreadsheet: For each column: does it hold identifiers, numbers, or something else? For identifiers: what is their scope and how are they information-bearing? For numbers: ordinal, cardinal or measurement? For measurements: what is the unit of measure and how are they connected? For rows: do they ever have to be in a particular order? For cells: what does empty mean? For values: are there special values treated differently? Represent the spreadsheet on a diagram: For entities: what must appear in the spreadsheet? For roles: which ones cannot repeat? For combinations of roles: which combinations cannot repeat? Real-world data: Representative sample? Legal terms?

51 Templates Breakdown Everything in blue on the diagram contributes to a template definition: The chain of boxes illustrating the relation The three role lines connecting the boxes to the entity ovals The role name haspossessor and its inverse participatesaspossessor The role name hasscale and its inverse participatesasscale The role name hasvalue The index of each role (1, 2, 3), left to right in the chain of boxes The uniqueness constraint line over role 1 The mandatory constraint dot on role 1 The description transponder has maximum operational pressure The entity oval around the relation and description. What s not part of the template? Everything in black: The solid line ovals representing named entities The dashed line oval representing the real number data type conceptual extent of a template definition by example What s not part of Object Role Modelling (ORM)? The dashed part of the uniqueness constraint (we add to be clearer). The individual role names and their inverses (not needed for ORM). The column labels A & C (we add to tie back to the spreadsheet)

52 Summary Reference data: makes data interchangeable stamps the meaning on data allows re-use of workflows allows re-purposing of data ISO Templates: describe spreadsheet-like data with rules and names Template definitions can make good reference data. Describing Information Flows Start with real-world data. Uncover implicit information. Describe it in sentences. Diagram it with ORM. Release it with sample data. The first step to reference data. Thanks!